The ‘Gambrel or Mansard?’ Problem

This post has been revised umpteen times, as the topic to be covered is somewhat complicated…

English is a language, like all languages, that is in a continual process of change. Some of these changes come about by words being imported/borrowed (and often mangled) from other languages, some from new ideas/technologies that need new descriptors, and some, dare I say most, changes result from accident or pure ignorance in the part of the speakers involved, and the repetition of such mistakes made by others. I’ll count myself among those who are ignorant in that regard, and admit I have a lot to learn yet about the English language.

The ways language can change is a fascinating topic in and of itself, and outside the scope of this blog for the most part. That said, I find I often have strong feelings about certain words and am attached to their meanings, and when I learn that I have been mispronouncing a word, or using it incorrectly, I am always happy to take such lessons forward with me. Some revisions are however largely futile, since other speakers of your language may no longer understand your speech if you make unconventional changes to it, and some changes happened so far in the past that it is all but pointless to hope to revert. Now, how far back is too far back is a question I’ll leave to others.

One case in point is the word ‘apron’. This word came into the language around between 1275 and 1325. The Middle English origin of this word is napron, which comes from the Middle French word naperon, a dimunitive form of the word nape, meaning ‘tablecloth’. Nape stems from the Latin mappa, which means ‘napkin’. In a curious process called juncture loss or rebracketing (<–a link) the 'n' can migrate back and forth between words beginning with vowels. Thus what was once 'a napron' became 'an apron'. That crafty little 'n' – ya gotta keep an eye on it like a three year old child or it can get into all sorts of trouble! Another English word that suffered from juncture loss is 'newt'. Originally the word was ewt, which etymologically relates to the word ‘eft’ (an eft means an immature newt), so ‘an ewt’ became ‘a newt’ and the shift stuck, this time the ‘n’ taking a sideways trip in the other direction. Same juncture loss process for an adder (the snake), which once was ‘a nadder’.

Another example of a word usage that has stuck, for completely different reasons, and the subject for my post today, is the word gambrel. I’m not sure how well known this word might be to readers outside the US, Canada, Australia, and to some extent England, but it is a word typically used to describe a building with a roof that looks like this:

While to many readers, that might well seem to end the discussion on the topic, I am not so inclined to stop, personally, because from the research I have done so far it would appear that the word gambrel has become, due to some accident of history, mis-applied – and after all I’ve got a blog to write this seems like a worthy topic to investigate!

‘Gambrel’ is a Norman English word, sometimes spelled gamerel, gamrel, gambril and gameral and having the meaning a crooked or hooked stick. A ‘Gambrel’ defined in a modern dictionary as a stick or piece of timber used to spread open and hang a slaughtered animal by its hind legs. Here’s a picture of a modern form of that device:

Here’s a picture of an old patent application for an adjustable gambrel:

As you can see, the form of the gambrel is that of a bar with hooks on the end, and in concert with the suspension chains, takes the form of a triangulated arrangement.

In action, the gambrel serves to suspend the slaughtered animal and facilitates the removal of the hide and meat from the leg areas, among other tasks:

Notice how, in the above example at least, the triangular gambrel angles outward in use. I may be reading something into nothing however.

The word gambrel also gets applied to the frame which an animal’s carcass is hung off of, as in this example:

The word ‘gambrel’ has another definition; it is a term for the joint in the upper part of a horse’s hind leg, the hock:

As an anatomic structure, the gambrel of the horse’s leg, when you peel the skin and surrounding tissues back, is actually a projecting spur of bone, and looks like this:

If one considers only the definition of ‘gambrel’ as a horse’s hock, and ignores the other definition of hooked stick, and ignores the fact that a ‘gambrel’ is a horizontal speading bar, then one could simply see the bent portion of the horse’s leg, the hock, and associate it to the roof that looks like this:

If you’re going to do that though, then one has to take a somewhat peculiar view of the horse’s leg arrangement, since the form of the above roof is only clearly seen at the gable ends. By that I mean: if that roof shape were to be akin to a horse’s back leg, then taking both roof sides together would give a view of a horse with an impossible akimbo stance. The stance, if were to be anything, would be that of the bow-legged cowboy who rides the horse! And there is no horizontal bar evident, nor does the roof lean outwards like the slaughterhouse gambrel shown in use above. So, I thought the association of the older meanings of ‘gambrel’ with the modern didn’t quite accord somehow, at least not to me.

Digging into the etymology of the word ‘gambrel’ a little further, I find that it comes from the Old Norman french gamberel, which is akin to the modern French jambier (meaning legging) and jambe, meaning ‘leg‘. Tracing further back, we have Middle French gambade, or gambolde, which means a leap or a spring. In modern English, we have the perhaps seldom-used word ‘gambado‘, which means the spring or leap of a horse. These words, along with ‘gambit’ and ‘gambol’, trace back to the Latin gamba meaning ‘leg’. Gamba in turn comes from the Greek, kampe, meaning ‘bend’.

While knowing the etymology back to the ancient Greeks is interesting in its own way, it must be said that the meanings of the words that derived from kampe and gamba are only of relevance in terms of what they meant at the time the epithet ‘gambrel’ was applied to a roof form. Since the slaughterhouse equipment/hooked stick/springing leg meanings of ‘gambrel’ were certainly in place before the use of the word as a roof name, we can presume, I think, that the roof shape called ‘gambrel’ was so-named for either its perceived physical similarity to, or connection semantically with, the slaughterhouse gear or the springing leg or spreading bar motif.

Returning to those word roots and their derivatives again, one comes to the sense of the word ‘gambrel’ as having a meaning of a leg bent back, a leg ready to spring. The gambrel stick bends the legs of the animal back and apart. The horse’s hock shows a bone that looks bent back, forming a crooked line. Again, although the gambrel roof as commonly described does indeed have a bend in it, there is nothing to suggest leaping or springing in the form. Nothing juts out, or bends back, in the roof now called ‘gambrel’ in N. America.

The intrigue of this question as to the origins of the word ‘gambrel’ was furthered for me personally when I took a class in French carpentry drawing a few years back, taught by French Compagnon Boris Noel. He was showing some architectural slides, and at one point I asked him what the French call the following roof form:

He replied, “Mansard”. No surprise there. Then I asked him what the French called the same roof form, but in a two-sided version, with gable ends. His answer: “Mansard”(!).

That was an ‘a-ha!’ moment for me.

So, while the French seem quite happy to call both roof forms ‘Mansard’, in North American usage (I would include Australia as well) the term ‘Mansard’ is applied only to roofs in which the doubled roof planes extend around all four sides of the building.

The French idea of a Mansard roof, as either 2- or 4-sided, would appear to be the definition used in other parts of continental Europe, as this clipping from a Turkish site indicates – both types of roof at the top of the picture are termed ‘Mansard’:

I also checked various German sites (one link), and found nothing for ‘gambrel’, but under ‘Mansard-dach‘ one finds illustrations of both 2- and 4-sided roofs with double pitches. While the French and Germans often have trouble agreeing on a wide variety of matters they seem to concur with their roof terms at least.

The word ‘gambrel’ in application to the roof form then came really from American vernacular usage – and likely confounding of terms. I think the English borrowed it later from this side of the pond to describe that bent-plane roof form roof form. I can’t prove that as of this moment, but it’s the theory I’m moving forward with.

Here’s another outrageous theory to consider: prior to permanent European settlement in America, Dutch mariners and traders had visited or settled the area of south east Asia now called Indonesia. It was there that they saw dwellings with a roof style where the end of a roof started as a hip and finished as a gable end at the ridge. The gable end was in fact an opening in the roof to allow smoke to dissipate from the cooking fires. This design of roof was brought back to Europe, where it saw minimal adoption – however, in the American Colonies the situation for building was a little less restricted by traditional forms, and the Indonesian roof form was adapted to local conditions in some places. This likely happened in the Dutch colonies along the Hudson, who also built roofs in the Mansard form, as it is a very practical kind of roof.

The roof style where the end of the roof starts in a hip and finishes in a gable – the hipped gable – is still in existence today in Indonesian rural communities:

A view from the side of an Indonesian hipped gable roof:

Notice in the side view the strong resemblance to the horse’s leg, the bend at the gablet akin to the hock, and how the gable at the top also seems to spring forward from the line of the hip. Notice if you took both sides of the hipped gable together, the lines of the hip ridges through the gable to the peak, the lines would not be so different than that of a horse’s rear legs as it was crouching to spring. The line of the gablet jutting out is similar to the line of the bone in the horse’s leg which produces the hock form. Also notice how, like the gambrel stick and chain assembly used to slaughter an animal, pictured earlier, the triangular gable pediment leans forward from the end hipped roof pitch. Also, considering the gambrel stick again, the bottom of the gablet on a hipped gable is defined by a horizontal bar, and forms a triangle. The hipped gable seems to tie far more clearly to the word ‘gambrel’ in it’s other senses. I think that somewhere along the line, the terms for the different Dutch roof forms got confused in the New World.

The word gambrill was part of the Dutch language in 1601 according to one account I read, though I have not been able to confirm this as of yet as it was no citation for the comment. Perhaps any Dutch readers of this blog who have access to libraries with old dictionaries might be able to assist me in this matter?

Various references are found in the original colonies in America about gambrel roofs including: 1737 Old Times, New England “One Tenement two stories upright with a gambering roof.”; 1765 Massachusetts Gazette “A large building with two upright stories and a Gambrel Roof. Sometimes with the long sloping roof of Massachusetts oftener with the quaint gambrel of Rhode Island”; 1779 “The gambrel ruft house”. 1824 “In a Gambrel roof’d home”; 1858 “a small farm with a modest gambrel roofed one story cottage”.

Those are references from New England mind you, and it is said that the ‘gambrel’ originated in The Hudson River Valley with the Dutch settlements.

In the Dictionary of Americanisms by John Russel Bartlett, 1848, pg. 153:

This dictionary is available online from Google books and has many interesting descriptions of all sorts of terms, many of which I had no idea originated in the US.

The question at this point is what John Bartlett meant by ‘hipped’? If he meant it in the sense we consider ‘hipped roof’ today, then this roof is not anything like the one we call ‘gambrel’ today.

The French or Mansard roof is attributed to the French Architect Nicolas Francois Mansart, 1598 – 1666, probably not something he invented but but certainly a roof style used by him. In its basic form it consisted of a King post truss (more correctly: King Piece Truss) on top of a Queen Post Truss. This provided usable roof space as additional accommodation, and possibly was intended, depending upon which account one reads, to take advantage of tax regulations at the time, or for use in areas where bricks and stone were expensive (as bricks and stone formed the walls).

The basic Mansard roof with gable ends could be called a single or 2-sided Mansard roof with the roof having two different pitches, the lower (or pitch from the eave) being steeper than the upper pitch connecting with the ridge. A Mansard Roof which has hip ends is also called a curb roof (except by the French and the rest of continental Europe, who stick to their guns and call it a ‘Mansard’) where the upper pole plates become a curb. The word ‘curb’ in this sense means to control or restrain, or perhaps enclosing framework or border. A synonym for the curb of the roof would be purlin plate. On some roofs, the curb is exposed and covered with fascia, marking a clear change in roof planes:

The roof shape of the Mansard was varied and pitch proportions were modified over time to accommodate dormer windows, and sometimes curved ends were formed at the eave to reduce snow slippage. Some roofs kept the double roof planes but made one or both planes curvilinear. Looking in old carpentry texts reveals numerous proportioning methods for the Mansard roof.

Now, as a counterpoint, one can imagine that a double pitch roof in which the lower edge of the lower plane were curled upward somewhat – hardly common but a form the roof can take – would have a strong resemblance to the hooked stick meaning of gambrel, so that is one possibility as to how the roof form names may have got confounded:

In the 1919 publication The Colonial Architecture of Salem, by Frank Cousins and Phil Riley, they devote an entire chapter to ‘Gambrel Roofed Houses’. There, they describe a gambrel roof as “an evolution of the seventeenth century Mansard roof“. Later, they describe how the roof form was adopted:

I wonder who these well-traveled ‘American Builders’ were who visited Paris and were impressed by the exposed cross-sections of Mansard? It’s an unattributed passage without footnote or reference. I also wonder about the account in regards to many of the buildings described in that chapter, as they have build dates in the late 1600’s and early 1700’s, while the word ‘gambrel’ didn’t come into use to describe a roof until 1735 at the earliest according to my dictionary. The locals and the builders had to be calling that roof something prior to 1735, and I wonder what the name was?

While an online dictionary search for the term ‘gambrel’ will lead to the current usual definition in most cases, in some architectural Dictionaries, like Dictionary of Architectural and Building Technology 4th Edition (by Henry J.Cowan and Peter R.Smith, published in the US by Spon Press 2004. Library of Congress ISBN 0-415-31234-5) you will find a different definition, to wit: “gambrel roof: A sloping roof similar to a HIP ROOF, but with the addition of small gables part-way up the end sloping portions”.

Another book I have, The Complete dictionary of Wood, by Thomas Corkhill (1982, Scarborough Books, ISBN 0-8128-6142-6) also shows the gambrel roof illustrated, on pg. 211:

With the picture, comes the following definition:

“Gambrel or Gambril roof. The end of a roof partly hipped and partly gabled.“

Or consider this English site, looking at buildings, and their depiction of roof forms:

It seems not all the English are in agreement about the word ‘gambrel’. Notice that the gambrel they depict matches that of Corkhill. Their written definition for ‘gambrel’?:

“A hipped roof which turns to a gablet at the ridge.“

Notice at the right of the above picture the various forms of gables depicted. Curiously, if you use the term ‘Dutch Gable’ in Australia or the US, the image normally ascribed to that term is this:

It would appear from what I came across while searching, that in Australia, what I call the hipped gable roof is a very popular form for detached open air car garages.

Frankly, the term ‘Dutch gable’ is only helpful in regards to the fact that it associates the hipped gable form to the Dutch. In truth, as mentioned earlier, the form is not actually Dutch, but Indonesian. As far as what a REAL Dutch gable looks like, in terms of buildings located in Holland, well, the form is a wee bit different. Of course, the Dutch don’t call it a ‘Dutch gable’ – one form of real Dutch gable, for example, is termed a klokgevel (‘bell-shaped gable’):

The ‘Dutch gable’, gevel, in fact is generally a type of facade on the short end of a building. Here’s an excellent link to view some of the various types of real Dutch gables (<– link), and here's a look at the structure of one of those classic Dutch buildings:
Speaking of the Dutch, thanks to the wonders of translation software and several hours of searching, I have a had a good look around several Dutch architectural websites trying to research the gambrel/Mansard conundrum. The term ‘gambrel’ is nowhere to be found (nor can I find a translation for in in a Dutch dictionary), however on one site which details a wide variety of roof shapes, we find the following entry for Mansard-dak (‘dak’ like the German ‘dach’, meaning ‘roof’):

“A Mansard is a gable or hipped roof, each of which was nodded. The under surfaces are steeper than the upper surfaces which creates more room on the top floor.“

Note: -A gabled OR a hipped roof-

So, again it would appear that most of continental Europe has the term ‘Mansard’ to describe 2 or 4 sided roofs of double pitch, and it would appear that they can live with that situation in a most untroubled manner, while we here in North America, and to an extent in Britain seem to prefer to have separate names for roofs with double roof planes (that are bowed out), even though one of the names may once have referred to a hipped gable roof. It’s quite a tangled nest of information to wade through and get to the bottom of, but that is what I intend to do as I continue my inquiry into this topic. I will have a follow-up post in this thread at some point in the future. I am in contact with a few different people who may be able to point the way, and I intend to do some research in the next month or so at the Boston Athaneum and then the Library of Congress in Washington D.C., in terms of looking at architectural and builders dictionaries from the 1750~1850 to see what they might have to say.

Thanks for visiting! I must confess my post was long and rambling today, and by no means cohesive or ever proving anything – yet! – so my apologies. Your learned comments or insights are most welcome.